face-to-face group

– see group(s).

face work

– see Erving Goffman.

factor analysis

One of the most widely used, and misused, of the complex multivariate statistics that have become more accessible since the spread of computing power, this can reduce a larger number of measured variables into a smaller number of latent variables, or “factors.” It is thus a “data reduction” technique, aimed at simplifying data while retaining its important features.

Factor analyses take as their input a number of different variables, usually all measuring similar related constructs (such as items in a standard personality questionnaire). The correlations between these measures are computed, and the number of dimensions (in a multi-dimensional space) that one needs to extract to describe the important variance, while screening out the error variance, is estimated. These factors are then constructed, and rotated to facilitate their interpretation. Finally, each case can be given a score on the newly created factors, for instance to describe respondents’ personality along each dimension of the model.

Factor analysis was critical to the conceptualization and development of research into intelligence and personality in early and mid-twentiethcentury psychology. For instance, R. B. Cattell, in The Scientific Analysis of Personality (1965), started by extracting all of the words in the English language to describe personality. Even after removing synonyms, there were still thousands of words. So Cattell used factor analysis to reduce this down to a list of fourteen personality scales, which became a standard model for many years in personnel selection and social research. H. J. Eysenck, in The Scientific Study of Personality (1982), went a stage further, and produced a model with just two dimensions, extraversion–introversion and neurotic– stable (and later a third dimension, psychoticism).

The initial appeal of factor analysis is that it would provide a scientific basis for answering some fundamental questions, such as how many dimensions there are to human personality or intelligence. Unfortunately, this promise to provide a scientific objectivity has not materialized, and in many cases different researchers, each using factor analyses, have arrived at very different conclusions.